CFE 2015: Start Registration
View Submission - CFE
A1695
Topic: Title: Efficient multipowers Authors:  Aleksey Kolokolov - Alliance Manchester Business School (United Kingdom) [presenting]
Roberto Reno - University of Siena (Italy)
Abstract: We show that the asymptotic variance of realized multipower estimators can be reduced considerably by combining different estimators and minimizing with respect to powers. After providing a general criterion for efficiency, we propose a new quarticity estimator which, thanks to the smaller variance, has the smallest mean square error among popular competitors on realistic simulations of financial prices. The application on US stocks provides further evidence of the superiority of the new estimator with respect to competitors, and in particular it shows that i) consistently with the theory, the distribution of the newly proposed quarticity estimator has smaller variance and thinner tails; ii) the new quarticity estimator does not suffer of the distortions due to the presence of jumps displayed by competitors; iii) using the new quarticity estimator instead of realized multipower variation in a standard test for the presence of jumps in asset prices (based on the difference between realized variance and bipower variation) improves the size and the power of the test.